[llvm] [RISCV][VLOPT] Add floating point widening and narrowing convert support (PR #122353)

[Philip Reames via llvm-commits]

https://github.com/preames approved this pull request.

LGTM - but please change the submit title to make it clear this is only the bf16 converts.  As written the description sounds much more generic.

https://github.com/llvm/llvm-project/pull/122353